The importance of materials properties in high-temperature processes

Ken MILLS1,2, Muxing GUO3

1 Dept. Materials, Imperial College, London SW& 2AZ,

2 University of Manchester 3 Dept. Metallurgy & Materials Eng. KU Leuven,

Physical properties &high-temperature processes

• Phys. Chem. Properties- useful in solving problems with process control and product quality.

• First law of high temperatures • “At high temperatures everything reacts with

everything else” • Second Law: • “They react bloody quickly and situation worsens

rapidly as the temperature increases”. • Many problems in high temp. processes- disaster

never far away

Example- variable weld penetration

TIG/GTA welding-robotic process-many welds

-set up optimum welding- Deep weld penetration - Change steel- meets

spec. -→shallow welds - Forces acting in pool

Effect of tiny differences in S

1400

1600

1800

2000

1500 1600 1700 1800 1900

Temperature, T/K

Surf

ace

tens

ion,

mN

m-1

• Hot liquid → melt back <50 ppm S –flow outward-

melt back-shallow weld >50 ppm S –flow inward &

down → deep weld

Surface tension >50steels 50ppm S →25% ↓ in γ d γ/ dT → ─ve to + ve

30 ppm S

80 ppm S

S< 50 ppm S> 50ppm

Slag properties and structure

Problem caused by ppm differences in S

Math models improved → insights to mechanisms

-Slags- are ionic 4Al steel+3SiO2 slag=3Si steel+ 2Al2O3 slag

4Al + 3Si4+ = 3Si + 4Al3+. Slag props –depend on slag

structure

Silicates- building block Si- 4O tetrahedron with O

joining 2 terahedra Cations,Na+ - break bonds

BOs, NBOs and free O’s

Alumino-silicates

Add Al2O3-network breakers↓

→ polymerisation ↑. (4)Al High Al2O3→some (5)Al, (6)Al NBO/T=2 (ΣXMO + ΣXM2O-XAl2O3 )/

(XSiO2+ 2 XAl2O3) Q=( 4- NBO/T) =measure of

polymerisation TiO2 behaves like SiO2? Fe2O3,Cr2O3-behave as Al2O3 ?

Al Na+

Charge-balancing

Si4+; Al3+; →(NaAl)4+; Na+ cannot act as network-breaker- polymerisation ↑

Effect of structure on slag properties

Polymerisation (Q) has great effect on props

Q↑ as %SiO2 ↑ as CaO%↓ Alumino –silicates May be Al-O bond weaker

than Si-O bond

-2

2

6

10

14

18

-1 0 1 2 3 4Q

ln η

1900

K, (

dPas

)

-2

0

2

4

6

8

-1 0 1 2 3 4Q

ln η

1900

K, (

dPas

)

Factors affecting properties

Slags- crystals or glasses Crystals-atoms fixed pos’n Glasses-disorder- forms

scl- high Sconfig; Nature cations→structure Bond strength- (z/ r2) (z/ r2)↑→ ( Qn→Qn-1+Qn+1) ↑ → Coord no ((6) N) ↑ Size: elect cond. Diffusion No. (n):Na2O n=2;CaO,n=1

Effect of structure- props 1. Viscosity(η) El cond

(κ)/ Res(R)Diff coeff(D) Liq: η = AA exp(BA/T) Scl: η = AV exp(BV/(T-To)) 1.Resistance cation movement 2. Mobility of cations size Equivalent: η; R; (1/D)

ln η1900; ln R1900;Bη;BR as f (Q)

• )

-2

2

6

10

14

18

-1 0 1 2 3 4Q

ln η

1900

K, (

dPas

)

0

20

40

60

0 1 2 3 4Q

B/1

000

-3-113

579

11

0 1 2 3 4Q

ln R

1900

K

0

10

20

30

0 1 2 3 4Q

BR

Density (ρ),Thermal expansion coeff (α,β) Surface and Interfacial tension (γsl, γmsl)

• α ↑ as Q↓ and (z/r2) ↓ • αglass ↑sharply at Tg-

• glass→ scl: • Tg –Tliq: αscl ≈ 3 αcrys: ρ slightly affected by structure ρcrys>ρglass- better packing

γsl, γmsl surface not bulk property

Surfactants- in surface layer eg. S,O in steel

slags, B2O3, CaF2,K2O, Na2O γmsl = γm+ γsl- 2φ( γm γsl)0.5 γm ≈4 γsl : γm most important - S, O contents in metal

0

2000

4000

6000

8000

200 400 600 800 1000Temperature, T/K

ΔL/L

293 p

pm

02468

1012141618

0 1 2 3 4Q

105 α

Thermal conductivity (k) thermal diffusivity (a)

00.20.40.60.8

11.21.41.61.8

2

0 200 400 600 800 1000 1200 1400

Temperature, T/oC

The

rmal

con

duct

ivity

,Wm-1

K-1

00.20.40.60.8

11.21.41.61.8

2

0 200 400 600 800 1000 1200 1400 1600

Temperature, T/K

The

rmal

con

duct

ivity

,Wm

-1K

-1

Heat transfer over slag film -complex-cond’n,rad’n, conv Solid: keff= klat+ kR. kR=16σn2T3/ 3α* liq glass: kR=10klat. kR↓ by (i) FeO add’n (ii) cryst kcrys≈ 2kglass

2 methods: LP and THW kLP ≈ 5-10 kTHW. Problem un-resolved

LP

THW

THW

LP

Manipulation of properties for process optimisation-liquidus temp (Tliq)

1. Liquidus temp (Tliq) Freeze linings- slag attack

of refractory – Lurgi -Slag coating of refractory Slag splashing- BOS-LD Pig Fe+O2→FeO; Si→SiO2 CaO added → slag Great turbulence → Extend refr life and

minimise “downtime” Optim. slag:13%FeO:8%MgO

Forms MgO.Fe2O3- high mp- >50,000 heats without re-lining

Manipulation of slag Tliq, viscosity(η)

Drainage rate- (1/ η) Coal gasification ash →↑ η slag CaO added →↓ η Low melting regions CaO- SiO2 : XCaO = 0.4-0.6 CaO- Al2O3 :XCaO=(0.5-0.7) Fixed amount CaO-

disasters Ladle glazes→inclusions

-2

2

6

10

14

18

-1 0 1 2 3 4Q

ln η

1900

K, (

dPas

)

Manipulation of surface (γsl) and interfacial tension (γmsl)

Slag & metal entrainment -Loss of precious metals -Inclusions in CC steel

Velocitymetal > Velocityslag . mentrap=1.06x10-7 (ηsl) -0.255. (γmsl) -2.18

Min by incr ηsl and γmsl.

Manipulation of surface (γsl) and interfacial tension (γmsl)

Math Model of CC process Predicts standing wave

,vortices, “necking and detachment”

γmsl = γm+ γsl- 2φ( γm γsl)0.5 γmsl ↑ as S, O in metal ↓

B2O3, K2O, Na2O ↓

Other problems involving surface and interfacial tension

1.Removal of inclusions By Ar bubbling 2. Scale formation Na2O, K2O cause slag to

stick to metal-reacts with FeO→FeSiO3 (scale)

3. Separation of metal droplets

Fe3O4

FeO

steel

Manipulation of heat transfer via freeze lining

Heat transfer-very complex Cracking and stickers Heat flux, q – ds, kR, kcryst ≈2kglass ; qcryst< qglass. Cryst ↑→ kR ↓ Heat flux manipulated by 1. ds →Tsol ; ds ↑ as Tsol ↑ 2. % crystals in slag film

Harvesting enthalpy in slag

Steelmaking Energy ca. 30% total costs BOS slag –granulated –

water spray on molten slag→ flame

H2- decomposition of H2O Harvest H2. Problem – O2 + H2 mixture Tata- reseach programme-

estimated saving- 30 million $ p.a for average steelworks

Produced >70%H2+N2+CO2

H2 could be used for energy or as reductant

Pilot –ferrochrome plant

Discussion

• Vast range slag compositions high temp processes • Compositions may change on daily basis • Need for models to calculate props from chem comp. • Models should include structural characteristics • Can represent structure in terms of BO, NBO or Q or

thermodynamic functions • Challenge will be in representing certain structural

features eg. Ti4+ prefers company of otherTi4+ to Si4+

Conclusions

1. Slag properties proved very useful for solving process problems

2. Mathematical models can give insights into mechanisms and help solve problems.

3. Knowledge of factors affecting physical properties allows one to manipulate processes- including valorisation

4. Vast range of slag compositions in pyro-metallurgy - will need good models to estimate properties from chemical comp’n- where structure comes in.